Skip to main content

reflect

Private reliance mirror. Summarizes how the user worked with AI over the last week (task types, coaching skipped, one-way doors, how well lessons transfer, lessons that keep slipping) and suggests one concrete habit for next week.

ソース情報

リポジトリ
pipeshub-ai/sharp
ソースの最終更新活動
2026年10月16日 06:32
検出された SKILL.md の言語
英語
スター
4
フォーク
0

インストール方法

デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。

ソースファイルを確認

インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。

SKILL.md を表示中

SKILL.md
ソースの指示 · 読み取り専用プレビュー
name
reflect
description
Private reliance mirror. Summarizes how the user worked with AI over the last week (task types, coaching skipped, one-way doors, how well lessons transfer, lessons that keep slipping) and suggests one concrete habit for next week.
argument-hint
[days, default 7]
disable-model-invocation
true
allowed-tools
Bash(node ~/.sharp/bin/sharp.mjs *)
# /sharp:reflect Local stats for the last 7 days (never leave this machine). If `$ARGUMENTS` names a different number of days, re-run `node ~/.sharp/bin/sharp.mjs stats <days>`: ```! node ~/.sharp/bin/sharp.mjs stats 7 2>/dev/null || echo "{}" ``` Write a short, warm, honest reflection (≤ 12 lines): 1. **How you used AI:** the mix of task types (debug, design, learn, build). Note it if most work was handed off with nothing kept for later: no lessons saved, no reviews, coaching skipped most of the time. 2. **What's sticking:** recall rate, reliably applied lessons, and "leeches" (lessons that keep failing to transfer, which point to a concept worth relearning with /sharp:learn). 3. **Judgment:** irreversible actions that were stopped for a decision, decisions recorded, and any past decision now due for a look back: name it and say what was chosen in words, never just an option letter. 4. **One habit for next week:** a single, specific, small suggestion, e.g. "Before asking for a fix, write one line saying what you think is wrong." Not a list. Close with the question: *"This week, how often did you use AI to help you think, and how often so you didn't have to?"* Let them answer, or not. This is for the user only. Never frame it as a score, and never compare them to others. ## How to write - Plain words, for a smart newcomer. Define each technical term in a few words the first time you use it. - Short paragraphs. Finish every part you start; never end on a fragment. - Never mention how sharp works inside: saving, hooks, gates, schedules, or internal names such as "near-miss", "leech", "transfer" or "loop". - Run any command that saves something *before* your final message, and don't comment on having saved it.
GitHubで見る